Pattern Recognition Based on YIQ Colour Space with Simulated Annealing Algorithm and Optoelectronic Joint Transform Correlation

Abstract

For pattern recognition on various views of the interested colour object, we adopt the YIQ colour space when using simulated annealing algorithm to design the template matching function. Joint transform correlation is devoted for recognition of colour targets. Quantized reference functions are designed for the purpose of display on liquid crystal spatial light modulators. Each reference function is trained with true class images rotated in-plane at 2 degrees intervals between -14 degrees and 14 degrees. Numerical result shows that, generally, YIQ space outperforms conventional RGB space.

Distortion Invariant Object Recognition Based on Mach-Zehnder Joint Transform Correlation in YIQ Colour Space

“…I is deviations from orange-luminance to cyan-luminance and Q is deviations from purple-luminance to chartreuse-luminance. Moreover, the optoelectronic system is based on a MZJTC structure [8]. It is consisted of one laser, one spatial filter, one collimated lens (CL), 3 beam splitters (BSs), 3 polarizing beam splitters (PBSs), 3 Fourier lenses (FLs), 3 reflective liquid spatial light modulators (RLCSLMs), 3 charge coupled device (CCD) cameras, 1 electronic subtractor (ES) which is used for removing the zero-order term at the final output, and 1 computer for controlling the whole system.…”

“…The output contains the overlapping of each cross-correlation of the reference channel and the target channel. More detailed analysis of MZJTC can be found in the literature [6][7][8].…”

“…To deal with the DC term, Lu et al [4] adoped phase-shifting technique to design a nonzero-order JTC (N0JTC) and Li et al [6] used the joint transform power spectrum (JTPS) subtraction strategy to realize the N0JTC. The Mach-Zehnder JTC (MZJTC) [6][7][8] can remove the zero-order term in only one step directly without storing the Fourier spectra of both the reference and target images beforehand. Later, Chen et al [9,10] adopted constraint optimization based on Lagrangian method to yield a sharp correlation peak.…”

Optical Correlators for Cryptosystems and Image Recognition: A Review

“…The proposed method was validated by numerical simulations, that showed a significant improvement of the PCE (peak-to-correlation energy) in the YIQ color space as compared to the conventional one based on RGB (red-green-blue) [ 22 ]. Earlier research by the same group had proposed a simulated annealing algorithm for pattern recognition using the MZJTC correlator in the YIQ color space and had shown, by numerical simulations, a significant improvement of the PSR (peak-to-sidelobe ratio) in the YIQ color space as compared to the RGB one [ 91 ].…”

Structured light based optical pattern recognition:Principle and technical progress

“…并且随着空间光调制器性能的愈加稳定, 人们可以 摆脱全息法制作匹配滤波器的非实时识别, 采用光电 混合型实现准实时识别. 除了实时性的追求以外, 人 们还实现了对复杂图像识别, 比如人脸 [43][44][45][46] 、DNA序 列 [47] 、虹膜 [48] 等生物信息用于安全系统的识别以及三 维 [49][50][51][52] 、彩色物体 [53][54][55][56] 等其他维度的复杂识别. 最近 JTC的概念还被应用于编码解码领域 [57][58][59] .…”